Comments on: Standalone Migrations: Using Rails migrations in non Rails projects

By: Todd Huss

Todd Huss — Wed, 08 Jul 2009 22:45:13 +0000

Great suggestions Bryon, I’ve implemented all of your changes (and re-added the database_sample.yml file).

By: BryonB

BryonB — Wed, 08 Jul 2009 01:17:06 +0000

I wanted to be able to specify different environments in the database.yml file like I can in Rails, so I modifed the :ar_init task in the Rakefile as follows:

task :ar_init do
require ‘active_record’
require ‘yaml’
ENV['RAILS_ENV'] ||= ‘development’
config = YAML.load_file(APP_BASE + “/config/database.yml”)[ENV['RAILS_ENV']]
ActiveRecord::Base.establish_connection(config)
logger = Logger.new $stderr
logger.level = Logger::INFO
ActiveRecord::Base.logger = logger
end

Now, I can pass RAILS_ENV as a parameter to db:migrate, like I can with Rails migrations, and it will use the ‘development’ environment by default.

Also, adding the following line at the end of the migrate task, to invoke the db:schema:dump task causes a schema.rb file to be dumped at the end of each migration, like it does in Rails. This is handy if you want to be able to use db:schema:load to set up a new environment from the schema.rb file.

Rake::Task[ "db:schema:dump" ].execute

Cheers!

By: BryonB

BryonB — Tue, 07 Jul 2009 21:37:28 +0000

Todd, there doesn’t seem to be any config/database_sample.yml file in the GitHub project when I download it. Are there any differences between the database.yml file here and one in a standard rails project? Can you still define different environments (development, production, etc.)? If so, where do you define them?